Heat Stress and Public Health: A Critical Reviewcgch.lshtm.ac.uk/Heat Stress and Public Health A...
Transcript of Heat Stress and Public Health: A Critical Reviewcgch.lshtm.ac.uk/Heat Stress and Public Health A...
ANRV337-PU29-09 ARI 29 October 2007 17:22
RE V I E W
S
IN
AD V A
NC
E
Heat Stress and PublicHealth: A Critical ReviewR. Sari Kovats and Shakoor HajatPublic and Environmental Health Research Unit (PEHRU), London Schoolof Hygiene and Tropical Medicine, London WC1E 7HT, United Kingdom;email: [email protected], [email protected]
Annu. Rev. Public Health 2008. 29:9.1–9.15
The Annual Review of Public Health is online athttp://publhealth.annualreviews.org
This article’s doi:10.1146/annurev.publhealth.29.020907.090843
Copyright c© 2008 by Annual Reviews.All rights reserved
0163-7525/08/0421-0001$20.00
Key Words
heat waves, early warning, mortality
AbstractHeat is an environmental and occupational hazard. The preventionof deaths in the community caused by extreme high temperatures(heat waves) is now an issue of public health concern. The risk ofheat-related mortality increases with natural aging, but persons withparticular social and/or physical vulnerability are also at risk. Im-portant differences in vulnerability exist between populations, de-pending on climate, culture, infrastructure (housing), and other fac-tors. Public health measures include health promotion and heat wavewarning systems, but the effectiveness of acute measures in responseto heat waves has not yet been formally evaluated. Climate changewill increase the frequency and the intensity of heat waves, and arange of measures, including improvements to housing, manage-ment of chronic diseases, and institutional care of the elderly andthe vulnerable, will need to be developed to reduce health impacts.
9.1
First published online as a Review in Advance on November 21, 2007
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
HHWS: heat healthwarning system
NWS: nationalweather services
INTRODUCTION
Heat is a natural hazard, and much is knownabout the effects of high temperatures on thehuman body. Episodes of extreme tempera-ture can have significant impacts on health andpresent a challenge for public health and civilprotection services. Further, one of the morecertain impacts of future anthropogenic cli-mate change will be an increase in heat wavesin many populations, and such heat waves willbe more intense (35).
Human populations are acclimatized totheir local climates, in physiological, behav-ioral, and cultural terms. There are clear andabsolute limits to the amount of heat exposurean individual can tolerate. However, humancapacity to adapt to varied climates and en-vironments is considerable. Most homes havean indoor temperature of 63◦F to 87◦F, andpeople do not comfortably live in tempera-tures outside this range. The tolerance rangeof an individual is usually much less than thisand will narrow with age or illness.
This article reviews the current epidemi-ological information on the impacts of heatwaves and hot weather and the implicationsfor public health. This topic has become arapidly growing area of epidemiological re-search. Since the 2003 heat wave, most coun-tries in Western Europe have implementedsome public health measures for heat waves,mostly in the form of heat health warningsystems (HHWS). The relatively rapid devel-opment of these systems is a success of pub-lic health. However, although more is knownabout who is most vulnerable to heat waves,there is very limited evidence on the most ef-fective ways to prevent heat-related mortality,particularly in community settings.
The Effects of Heat on the Body
Healthy adult persons have efficient heat reg-ulatory mechanisms, which cope with in-creases in temperature up to a particularthreshold. The body can increase radiant,convective, and evaporative heat loss by va-sodilatation and perspiration (43). Experi-
mental data have been used to describe a widerange of thermal indices (more than 300) (79)and for setting important occupational andother standards to limit exposure to heat andthe associated health effects (77). The physio-logical effects of heat are reviewed extensivelyelsewhere, but there is a lack of evidence onheat tolerance in women, in the elderly, andin persons with chronic disease (see below).
High temperatures cause the clinical syn-dromes of heat stroke, heat exhaustion, heatsyncope, and heat cramps (42). Severe heatstroke occurs when the core body tempera-ture exceeds 103◦F and leads to multiple or-gan dysfunction. Heat stroke has a substantialcase-mortality ratio, and progression to deathcan be very rapid (within hours). In survivors,the permanent damage to organ systems (17)can cause severe functional impairment (16)and increase the risk of early mortality (97).
HEAT WAVES AND THEIRIMPACTS ON HEALTH
This review focuses on the prevention of heat-related impacts in community settings. Theheat wave in France in August 2003 caused14,802 deaths in a 20-day period (33). A ma-jor heat wave in Athens in 1987 was associ-ated with more than 2000 deaths (39). Otherwell-studied heat waves include several in theU.S. Midwest region, particularly the 1995Chicago event (46). What constitutes a heatwave event is loosely defined, and nationalweather services (NWS) have developed theirown definitions on a national or local basis.In practice, the term heat wave is applied to avery wide range of meteorological conditions,from moderate to severe. Heat waves in termsof a disaster or emergency, i.e., that involvesome aspect of the overwhelming of publicservices, are rare. Unfortunately, best practiceguidelines had not been developed in Europeor the United States until recently. Table 1lists the heat wave events in Europe that havebeen reported in the health literature. Majorheat wave events are also associated with otherhealth hazards such as air pollution episodes,
9.2 Kovats · Hajat
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
Table 1 Heat wave events and attributed mortality in Europe (adapted from Reference 52)
Heat wave event Attributable mortality (all cause) Baseline measure1976—London, UK 9.7% increase England and Wales and
15.4% Greater London31-day moving average of daily mortalityin same year
1981–Portugal 1906 excess deaths (all cause, all ages)in Portugal, 406 in Lisbon (month ofJuly)
Predicted values
1983—Rome, Italy 35% increase in deaths in July 83 in65+ age group
Compared with deaths in same month inprevious year
1987—Athens, Greece estimated excess mortality >2000 Time trend regression adjustedJuly 21–311991—Portugal 997 excess deaths Predicted valuesJuly 12–211995—London, UK 11.2% (768) in England and Wales,
23% (184) Greater London31-day moving average of daily mortalityin previous two years
July 30–August 31994—Netherlands 24.4% increase, 1057 (95% CI 913,
1201)31-day moving average of previous 2 years
July 19–312003—Italy, 3134 (15%) in all Italian capitals Deaths in same period in 2002June 1–August 152003—France 14802 (60%) Average of deaths for same period in years
2000 to 2002August 1–202003—Portugal 1854 (40%) Deaths in same period in 1997–2001August 1–312003—Spain 3166 (8%) Deaths in same period 1990–2002August 1–312003—Switzerland, 975 deaths (6.9%) Predicted values from Poisson regression
modelJune 1–August 31 (3 months)2003—Netherlands 1400 deaths Number of degrees above 72◦F multiplied
by the estimated number of excess deathsper degree (25–35 excess deaths)
June 1–August 232003—Baden-Wuertermburg,Germany
1410 deaths Calculations based on mortality of pastfive years
August 1–242003—Belgium 1297 deaths for age group older than 65 Average of deaths for same period in years
1985–2002
2003—England and Wales 2091 (17%). Mortality in Londonregion: 616 deaths (42% excess)
Average of deaths for same period in years1998 to 2002
wild fires, and water and electricity supply fail-ures, which also have implications for publichealth action.
The excess mortality attributed to a heatwave event is the short-term increase inthe numbers of deaths (Figure 1, see color
insert), a peak in mortality similar to that seenfor very severe pollution episodes. The esti-mated number of deaths therefore depend onthe definition of the heat episode. Reviewshave shown that the total impact of a heatwave event will be dependent on a number of
www.annualreviews.org • Heat Stress and Public Health 9.3
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
factors including heat wave magnitude, tim-ing in season, population experience of heatwave events, and public health responses (47).
The majority of heat wave studies haveconsidered impacts on mortality because dailydeaths data are generally readily available inhigh-income countries. Heat waves are alsoassociated with increases in emergency hos-pital admissions (36, 51, 84). During the Au-gust 2003 heat wave, an increase in admis-sions was reported in Spain (10) and in France,where many hospitals were overwhelmed (27,54, 96). Heat-related increases in emergencyadmissions are most apparent for particularoutcomes, including renal and respiratory dis-ease, particularly in the elderly (51). Highertemperatures are not associated with increasesin admissions for cardiovascular disease (51,76), although some effect is apparent in theUnited States (83). Health system factors,such as admission thresholds, may explainsome of this difference. Studies so far have in-dicated that increases in hospital admissionsduring heat waves are not as severe as thoseseen in mortality data. One reason for thismay be that people who die during heat wavesdie suddenly or do not reach the attention ofthe medical services. The latter hypothesis hasimplications for health protection measures.
METHODS OFCHARACTERIZING THETEMPERATURE-MORTAILTYRELATIONSHIP
Assessment of Heat Wave Effects
Heat waves have a greater impact on mor-tality than shown in the reported number ofdeaths or cases certified as classic heat illness.Figure 1 illustrates the peak in all-cause mor-tality associated with an acute heat episode.Heat episode analyses are traditionally usedto investigate health—usually mortality—during specific heat wave events (39, 48, 60,95, 100). These describe mortality counts orrates during a heat wave and are comparedwith a baseline usually derived from the same
time period in surrounding years. The periodof interest should differ from the compari-son period only with regard to a heat waveoccurrence, and so any general trends in thebaseline mortality series need to be avoided orallowed for. An alternative is to create the ex-pected values using a regression model withexplicit control for season and other time-varying confounders (see below).
Assessment of Heat Effects
Time-series regression. Regression mod-els of time-series data are often used to quan-tify the general heat-mortality relationshipobserved throughout the summer (8, 14, 41,101). The general principle of a time-seriesregression approach involves assessment ofany short-term associations between regularmeasurements of the health outcome and theexposure (e.g., daily death counts and dailytemperature over a number of years). Themethods used are closely related to those de-veloped for air pollution epidemiology (40,81). Unlike air pollution, however, the effectsof temperature on mortality cannot be as-sumed to follow a general linear form. In pop-ulations with a temperate climate, a generalU- or V-shaped relationship exists betweendaily mortality counts and temperature, withdeaths increasing as temperatures fall, but alsoas temperatures rise above population-specificthreshold values (14).
Multiple lag times are associated with tem-perature exposures, especially in relation tocold weather, but heat effects may also be de-layed by up to a few days. The effects of expo-sure on multiple days, including any negativerisks arising from mortality displacement (seebelow) may be modeled using a distributedlag curve (7, 8, 30, 99). The use of cross-basisfunctions has also recently been proposed toallow for flexible modeling of changes in co-efficients with changing lags (3). The key is-sue with the time-series design is the opti-mal control for confounding by season andother causes of fluctuations in mortality overtime.
9.4 Kovats · Hajat
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
Case-crossover study. If individual-levelmortality records are available, an increas-ingly common alternative to time-series re-gression is the case-crossover design, in whichthe day of death of each individual is consid-ered a case, and proximate days (e.g., +/−7days) are controls (89). The analytic approachis then akin to that for a matched case-controlstudy. This approach circumvents the need forseasonal control but can be subject to other bi-ases, especially in relation to the choice of con-trols (58). The method is being increasinglyrefined as knowledge accumulates about po-tential biases, and investigators have recentlyfavored a time-stratified design in which con-trol days are chosen from time strata a pri-ori (4, 57). Such designs are a special case ofthe general case-series approach (22). An im-portant consideration in both time-series andcase-crossover studies is serial correlation, thenonindependence of consecutive days.
Effect Modification
In addition to the approaches above, otherdesigns have also been used to assess effectmodification of the heat-mortality relation-ship. The case-only approach has been usedto quantify the modification effect of severalrisk factors but does not give an indication ofthe overall effect (2, 82). Several case-controlstudies have also been undertaken on heatwave events in Chicago (85) and Paris (94).These studies used live controls and there-fore may also be estimating factors that deter-mine the risk of death per se, rather than thedeterminants of a heat-related death. Strati-fying daily mortality series by subgroup canalso be used to investigate effect modification(26, 31). However, the number of subgroupsthat can be investigated using this method isnormally quite limited.
Mortality Displacement
A key consideration about the correct publichealth interpretation of the heat effect esti-mates is the extent to which heat deaths occur
AT: apparenttemperature
in already frail individuals whose death may behastened by the heat exposure by only a mat-ter of days—so-called short-term mortalitydisplacement or early harvesting. Studies us-ing heat episode analysis rarely consider post–heat wave periods to assess whether excesses inmortality during a heat wave are followed bysubsequent deficits in expected deaths, consis-tent with a harvesting process. Recent studiesusing penalized splines to model the mortalitypattern during the Paris (56) and Chicago (37)heat waves suggested little evidence of har-vesting during these very extreme heat waveevents. Assessment of a harvesting processduring a heat wave is complicated by the factthat any negative risk following an initial hotday may be masked by an increased risk fromfurther heat exposure from subsequent hotdays; these two opposing effects cannot beseparated. Time-series studies have demon-strated some degree of mortality displacementfollowing general heat-related deaths (7, 30).
Other Issues
Also of importance is the correct parameter-ization of the exposures measure. Mean tem-perature is commonly used, with control forconfounding with humidity. Combined in-dices of temperature and humidity, such asapparent temperature (AT), are also used as aconstruct that characterizes the physiologicalexperience better than just temperature alone(90). However, in a recent assessment, AT wasnot a better predictor of mortality than wasmean temperature in two of three Europeancities (29).
Although some major studies of temper-ature and mortality have not controlled forthe effects of air pollutants (7, 14), more re-cent work has suggested that heat effects arelikely to persist even after control for air pollu-tion, and weather risk assessments are best in-formed by analyses that account for PM10 andozone in particular (73). Heat episode analysesimplicitly control for the air pollution con-centrations because the baseline representsthe seasonal norm; however, there exists the
www.annualreviews.org • Heat Stress and Public Health 9.5
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
possibility of synergistic effects of air pollu-tion with heat, which may be more apparentduring extreme events (38).
DETERMINANTS OFHEAT-RELATED MORTALITYAND MORBIDITY
Public health interventions are often targetedat high-risk groups, and it is therefore impor-tant to identify those most vulnerable to dy-ing in a heat wave. Published epidemiologicalstudies are available for three main categoriesof exposure related to hot weather:
� individual heat wave events;� days with temperatures above a speci-
fied heat threshold, as described in time-series regression studies; and
� classic heat stroke related to hot weather(note that a significant proportion ofheat stroke is not related to heat wavedays).
Risk factors can also be categorized as in-trinsic (age, disability) and extrinsic (housing,behaviors); the latter vary according to loca-tion and adaptations to the local climate. Therisk factors can operate at many stages alongthe causal chain from high ambient temper-ature to death (Figure 2). Effective healthprotection measures have important effects
on the risk of heat-related mortality at thepopulation level and may modify the risks byage group (15). However, there is little pub-lished information as yet on the effectivenessof such measures, although it is clear that themortality response can change dramaticallyfrom one heat wave to the next (see below)(65).
Climate itself is an important determi-nant of population sensitivity to temperaturebecause it influences the level of acclima-tization in individuals. The threshold tem-peratures for heat-related mortality effectsare related to the local summer temperature.That is, the temperature above which mor-tality increases with increasing temperatureis higher in warmer climates compared withcooler climates. The slope of the temperature-mortality response, however, is rather het-erogeneous and in general not predicted bylatitude, as shown by comparisons of citieswithin the United States (14), Europe (66),and around the world (62).
Age and Aging
Vulnerability to heat in old age occurs be-cause of changes in the thermoregulatory sys-tem (23, 28, 93). Epidemiological studies ofheat-related mortality show a larger effect in
Heat Heat stressHeat illness
(clinical signs) Heat death
Factors affectingexposure
Factors affectingaccess to treatment
Factors affecting
sensitivity to a given heat
exposure
Figure 2Points along the causal chain from heat exposure to heat death.
9.6 Kovats · Hajat
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
the elderly; the risk increases with increasingage above ∼50 years old.
Children and babies also have limited abil-ity to thermoregulate. Children are also moreat risk of dehydration than are adults. Childdeaths from heat stroke occurred in Franceduring the heat waves in 2003 (78) and 2006(21). However, time-series and episode analy-ses indicate no excess in mortality in childrendue to heat waves [except the 1981 and 1991heat waves in Portugal (25, 74)].
Heat mortality risk varies by both age andsex. The majority of European studies haveshown that women are more at risk, in bothrelative and absolute terms, of dying in a heatwave. There maybe some physiological rea-sons for an increased risk in elderly women(9, 32), but social factors are also important.In Paris, the heat risk increased for unmarriedmen, but not unmarried women (11). In theUnited States, elderly men are more at riskin heat waves than are women, and this wasparticularly apparent in the Chicago events(85, 100). This vulnerability may be due tothe level of social isolation among elderly men(46). Evidence for the importance of socialcontact as protective for heat wave mortalityis based on case control studies conducted inthe United States (68, 85). Men are also moreat risk of heatstroke mortality because theyare more likely to be active in hot weather(13).
Clinical or PathophysiologicalFactors
In addition to the natural patterns of aging(or senescence) on homeostatic mechanisms,several medical conditions increase vulnera-bility to heat stress. Epidemiological studiesindicate that people with depression, cardio-vascular and cerebrovascular conditions (89),and diabetes (82) need to take extra care inhot weather. The literature on pathophysiol-ogy of heat is incomplete. Many deaths thatare attributed to heat are not caused by heatstroke nor are they even in persons that ex-hibit the clinical signs of heat stress. There
are likely several mechanisms by which a per-son may die during a heat wave because theenvironmental temperatures place extra strainon the cardiovascular system. The person be-comes dehydrated, displaying increased bloodviscosity and other physiological changes. Intheory, any illness that compromises ther-moregulation will increase the risk of heat-related death. If the exposure to heat is severeenough, even healthy people will succumb toheat stroke. Illnesses also compromise mobil-ity, awareness, and behavior. Therefore, someconditions, such as dementia and Parkinson’sdisease, are also important risk factors forheat mortality. In addition, a range of com-mon medications interfere with thermoregu-lation (anticholinergics interfere with sweat-ing, diuretics can cause dehydration, etc.)(17, 19, 44).
Living in Institutions
The heat wave in 2003 caused a severe impacton elderly persons in hospital and in residen-tial homes. The mortality rate doubled in the75+ age group for persons living in retirementhomes in France (24). In 2003, such institu-tions in France, Italy, and the United King-dom were generally without air conditioning.A study in the United Kingdom also found amuch higher heat risk in nursing home pa-tients and care home patients, with lowestrisk in persons living at home (31). The res-idents of institutions therefore represent animportant high-risk group for heat wave in-terventions. In France, the government hassince recommended that homes for the el-derly have at least one cool room (63). Inthe southern United States, there appears tobe little heat-related mortality in care homes,except when an air conditioning system fail-ure occurs (92). The risk of heat-related mor-tality is determined by level of disability (orfrailty) in the nursing home setting (5); how-ever, a study in France suggested that as med-ical care during the heat wave was directedtoward the most vulnerable patients, the less-frail patients made the largest contribution to
www.annualreviews.org • Heat Stress and Public Health 9.7
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
the excess mortality estimated during the heatwave (34).
Housing Characteristicsand Air Conditioning
People spend the most of their time indoors.There is relatively little evidence that certaintypes of housing (and thermal behavior) in-crease vulnerability. Brick houses (with a highthermal mass), top floor apartments with nothrough ventilation, and closed windows areassociated with an increased risk of mortalityduring a heat wave (67, 94).
U.S. studies indicate that air condition-ing is an important protective factor for heat-related mortality (85). In Europe, there is lit-tle domestic air conditioning, although thisis expected to change in the coming decades.Lack of air conditioning may explain the riskof heat stroke in poor urban elderly personsin some U.S. inner cities because the role ofhigh energy costs and the loss of income sup-port was an issue during the Chicago heatwave (46). Power failures are common dur-ing heat waves because of sudden increasesin electricity demand (20). Some investigatorssuggest that widespread use of air condition-ing may reduce physiological acclimatizationand can therefore make people more suscep-tible to heat waves, but the evidence is unclear(72).
Socioeconomic Factors
Deprivation, particularly in the inner city,is an important determinant of heat waveand heat stroke mortality risk in the UnitedStates, but the evidence is less clear for Euro-pean populations. The heat wave in Pheonixin 2006 was responsible for 13 heat strokedeaths, of which 11 were homeless people.Most studies of the 2003 heat wave in Eu-rope report little or no effect of deprivation onmortality risk (6, 31, 89), although one studyin Italy reported a higher risk in low-incomegroups (64). Individuals on low incomes aremore likely to have a chronic disease or other
medical risk factors, such as obesity or men-tal illness, and less adequate types of housing,which will all modify the risk of heat-relatedmortality.
Urban Heat Islands
Urban heat islands are a factor in many citiesand refer to the difference in temperaturesmeasured inside and outside the city (70, 71).Heat islands have been particularly importantin some heat wave events (53, 98). Heat is-lands are dynamic in time and space and aretherefore hard to quantify either spatially ortemporally for individual heat wave events.Although much is known about the factorsof the built environment that increase temper-atures, when estimating health effects withina city, confounding by housing type and so-cioeconomic factors becomes very important.Several studies show that mortality is moresensitive to heat in urban areas compared withrural and suburban areas. One reason for thiscould be that urban heat islands magnify nighttime temperatures. The importance of heat is-lands may vary, as cities in southern Europeare more adapted to heat than are those innorthern Europe (31, 87). A study in Spainfound that excess mortality during the August2003 heat wave was comparable in rural vil-lages and in the provincial capital (61).
IMPLICATIONS FORPUBLIC HEALTH
Many public health lessons were learned fol-lowing the August 2003 heat wave in westernEurope (86). The key problems identified bythe French government included the lack ofan intervention plan for heat waves and thelack of coordination between the social ser-vices and health agencies. In addition, veryfew care homes (“maison de retrait”) had airconditioning or other space cooling, and thesebuildings simply became too hot. However,the heat wave itself was so severe that it couldnot have been anticipated (59, 86). It is very
9.8 Kovats · Hajat
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
likely that anthropogenic climate change hasdoubled the likelihood of such events (91).
Many countries now have heat wave plansthat dictate a range of public health interven-tions. Health education about prevention andidentification of the first signs of heat stress isthe most important public health strategy, butsuch campaigns must be repeated at the begin-ning of every summer. HHWSs use meteoro-logical forecasts to reduce the impact of heatwaves on human health (49). The challengelies in determining at which point the weatherconditions become sufficiently hazardous tohuman health in a given population to warrantintervention (47, 50, 69, 88). Warnings issuedto the general public during a heat wave re-inforce the general heat-avoidance messages.The active components of HHWSs, such asidentifying and contacting high-risk individ-uals, vary from city to city. Intervention plansshould be best suited for local needs throughcoordination between the local health agen-cies, social services, voluntary agencies, andthe NWS.
The heat wave in 2006 in western Europewas associated with much less impact on mor-tality than that in 2003 (21). It is not straight-forward to compare directly the impacts ofheat waves in terms of numbers of deaths, ei-ther in different cities or in the same city overtime (15, 50). Fewer heat-related deaths oc-curred in Chicago during the 1999 heat wavecompared with the earlier 1995 event. Al-though some of this reduction in mortality wasattributed to the successful implementation ofprevention measures, such as the opening ofcooling centers (75), as well as an increase inair conditioning use, there would also havebeen a significant increase in the general levelof awareness heat waves impacts and publichealth messages.
Another important issue that has beenidentified is lack of health surveillance forheat wave mortality. In 2003, surveillance ofgeneral mortality on a daily basis was notavailable, and systems could not detect an in-crease in deaths in the elderly from nonspe-cific causes (1). Some countries consequently
Table 2 Recent trends, assessment of human influence on the trend, and projections for extreme weather events forwhich there is an observed late-twentieth-century trend [Fourth Assessment Report of the Intergovernmental Panel onClimate Change (35)]
Phenomenon and direction of trend
Likelihood that trendoccurred in late
twentieth century(typically post-1960)
Likelihood of ahuman contributionto observed trend
Likelihood of futuretrends based onprojections for
twenty-first centuryWarmer and fewer cold days and nightsover most land areas
Very likely Likely Virtually certain
Warmer and more frequent hot days andnights over most land areas
Very likely Likely (nights) Virtually certain
Warm spells/heat waves. Frequencyincreases over most land areas
Likely More likely than not Very likely
Heavy precipitation events. Frequency(or proportion of total rainfall fromheavy falls) increases over most areas
Likely More likely than not Very likely
Area affected by droughts increase Likely in many regionssince 1970s
More likely than not Likely
Intense tropical cyclone activity increases Likely in some regionssince 1970
More likely than not Likely
Increased incidence of extreme high sealevel (excludes tsunamis)
Likely More likely than not Likely
www.annualreviews.org • Heat Stress and Public Health 9.9
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
now use other indicators that are available ona 24-hour basis, such as a telephone advice linein the United Kingdom, to support decisionmaking for heat wave alerts (55).
CONCLUSION
Heat health protection strategies require fur-ther research to identify those that are trulyeffective in reducing mortality due to heatwaves. The limited epidemiological evidencesuggests that targeting high-risk groups maynot be the most effective strategy but that sus-tainable improvements in the health of thevulnerable elderly are required.
Global climate change is likely to be ac-companied by an increase in the frequencyand intensity of heat waves (Table 2). In thelong term, as the climate changes, populationsare likely to become less sensitive to temper-ature extremes owing to improvements in theunderlying health of the population (12, 18).Conversely, populations are aging, and thenumber of elderly people susceptible to tem-perature extremes will increase. There is also atrend toward more energy-intensive buildingsthat need to be artificially cooled (80). Thuswe must also build houses and cities that arecooler as well as more sustainable and energyefficient.
SUMMARY POINTS
1. A range of epidemiological study designs are used to quantify the effect of temperatureon mortality and explore effect modification, including time-series regression; episodeanalyses; and case-control, case-only, and case-crossover methods.
2. Mortality due to heat waves is most pronounced among the elderly, but other groupsare also at risk, including adults with chronic disease and children. The epidemiologi-cal evidence base needs to be reviewed regularly to determine the advice on targetingcertain groups with prevention strategies.
3. A range of measures, including health advice and weather-based alerts, are used toprevent heat wave deaths, but there is no clear evidence about the most effectivemeasures in community settings, particularly in targeting the very vulnerable.
4. Climate change will increase the frequency and the intensity of heat waves, and a rangeof measures—including improvements to housing, management of chronic diseases,and institutional care of the elderly and the vulnerable—will need to be developed toreduce future impacts of heat.
DISCLOSURE STATEMENT
The authors are not aware of any biases that might be perceived as affecting the objectivity ofthis review.
ACKNOWLEDGEMENTS
The authors thank Bettina Menne and Franziska Matthies, and all participants in the EU-ROHEAT network. The authors’ work was funded in part by the European Commission DGSANCO (agreement number 2004322).
LITERATURE CITED
1. Abenheim L. 2005. Lessons from the heat wave epidemic in France (Summer 2003). SeeRef. 45, pp. 161–66
9.10 Kovats · Hajat
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
2. Armstrong B. 2003. Fixed factors that modify the effects of time-varying factors: applyingthe case only approach. Epidemiology 14:467–72
3. Armstrong B. 2006. Models for the relationship between ambient temperature and dailymortality. Epidemiology 17:624–31
4. Basu R, Dominici F, Samet JM. 2005. Temperature and mortality among the elderly inthe United States: a comparison of epidemiological methods. Epidemiology 16:58–66
5. Belmin J, Auffray J-C, Berbezier C, Boirin P, Mercier S, et al. 2007. Level of dependency:a simple marker associated with mortality during the 2003 heatwave among French de-pendent elderly people living in the community or in institutions. Age Ageing 36:298–303
6. Borrell C, Marı-Dell’Olmo M, Rodrıguez-Sans M, Garcia-Olalla P, Cayla JA, et al. 2006.Socio-economic position and excess mortality during the heat wave of 2003 in Barcelona.Eur. J. Epidemiol. 21:633–40
7. Braga ALF, Zanobetti A, Schwartz J. 2001. The time course of weather-related deaths.Epidemiology 12:662–67
8. Braga ALF, Zanobetti A, Schwartz J. 2002. The effect of weather on respiratory andcardiovascular deaths in 12 US cities. Environ. Health Perspect. 110:859–63
9. Burse RL. 1979. Sex differences in human thermoregulatory response to heat and coldstress. Hum. Factors 21:687–99
10. Cajoto VI, Peromingo JA, Vicdeo GV, Leira JS, Frojan S. 2005. Health impact of 2003heat wave at Hospital de Riveira (A Coruna). An. Med. Int. 22:15–20
11. Canoui-Poitrine F, Cadot E, Spira A, Groupe Reg. Canicule. 2006. Excess deaths duringthe August 2003 heatwave in Paris, France. Rev. Epidemiol. Sante Publique 54:127–35
12. Carson C, Hajat S, Armstrong B, Wilkinson P. 2006. Declining vulnerability totemperature-related mortality in London over the twentieth century. Am. J. Epidemiol.164:77–84
13. Cent. Dis. Control (CDC). 2006. Heat-related deaths—United States, 1999–2003.MMWR 55:769–98
14. Curriero FC, Heiner KS, Samet JM, Zeger SL, Strug L, Patz JA. 2002. Temperature andmortality in 11 cities of the Eastern United States. Am. J. Epidemiol. 155:80–87
15. Delaroziere JC, Sanmarco JL. 2004. Excess mortality in people over 65 years old duringsummer heat waves in Marseille. Comparison before and after preventive campaign. PresseMed. 33:13–16
16. Dematte JE, O’Mara K, Buescher J, Whitney CG, Forsythe S, et al. 1998. Near-fatalheat stroke during the 1995 heat wave in Chicago. Ann. Intern. Med. 129:173–81
17. Dixit SN, Bushara KO, Brooks BR. 1997. Epidemic heat stroke in a midwest community:risk factors, neurological complications and sequelae. Wis. Med. J. 96:39–41
18. Donaldson GC, Keatinge WR, Nayha S. 2003. Changes in summer temperature and heat-related mortality since 1971 in North Carolina, South Finland, and Southeast England.Environ. Res. 91:1–7
19. Ellis F. 1976. Heat wave deaths and drugs affecting temperature regulation. Br. Med. J.2:474
20. Emerg. Manag. Aust. (EMA). 2002. EMA Disaster Events Data Tracking System (EMA-Track) http://www.ema.gov.au/
21. Empereur-Bissonnet P, Salines G, Berat B, Caillere N, Josseran L. 2006. Heatwavein France, July 2006: 112 excess deaths so far attributed to the heat. Eurosurveillance11:E060803.3
22. Farrington CP, Whitaker HJ. 2006. Semiparametric analysis of case series data. Appl.Stat. 55:553–94
www.annualreviews.org • Heat Stress and Public Health 9.11
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
23. Flynn A, McGreevy C, Mulkerrin EC. 2005. Why do older patients die in a heatwave?QJM 98:227–29
24. Fouillet A, Rey G, Laurent F, Pavillon G, Bellec S, et al. 2006. Excess mortality relatedto the August 2003 heat wave in France. Int. Arch. Occup. Environ. Health 80:16–24
25. Garcia AC, Nogueira PJ, Falcao JM. 1999. Onda de calor de Junho de 1981 em Portugal:efeitos na mortalidade. Rev. Port. Saude Publica 1:67–77
26. Gouveia N, Hajat S, Armstrong B. 2003. Socio-economic differentials in the temperature-mortality relationship in Sao Paulo, Brazil. Int. J. Epidemiol. 32:390–97
27. Gremy I, Lefranc A, Pepin P. 2004. Consequences sanitaire de la canicule d’aout 2003en Ile-de-France [Impact of the August 2003 heat wave: sanitary consequences in Ile-de-France]. Rev. Epidemiol. Sante Publique 52:93–98
28. Grundy E. 2006. Ageing and vulnerable elderly people: European perspectives. AgeingSoc. 26:105–34
29. Hajat S, Armstrong B, Baccini M, Biggeri A, Bisanti L, et al. 2006. Impact of hightemperatures on mortality: Is there an added “heat wave” effect? Epidemiology 17:632–38
30. Hajat S, Armstrong B, Gouveia N, Wilkinson P. 2005. Mortality displacement of heat-related deaths: a comparison of Delhi, Sao Paulo and London. Epidemiology 16:613–20
31. Hajat S, Kovats RS, Lachowycz K. 2007. Heat-related and cold-related deaths in Englandand Wales: Who is at risk? Occup. Environ. Med. 64:93–100
32. Havenith G. 2005. Temperature, heat balance, and climatic stress. See Ref. 45, pp. 70–8033. Hemon D, Jougla E. 2004. La canicule du mois d’aout 2003 en France. Rev. Epidemiol.
Sante Publique 52:3–534. Holstein J, Canoui-Poitrine F, Neumann A, Lepage E, Spira A. 2005. Were less disable
patients the most affected by 2003 heatwave in nursing homes in Paris, France? J. PublicHealth 27:359–65
35. Intergov. Panel Clim. Change (IPCC). 2007. Climate Change 2007: The Physical ScienceBasis. Contribution of Working Group I to the Fourth Assessment Report of the IntergovernmentalPanel on Climate Change. Cambridge, UK/New York: Cambridge Univ. Press
36. Johnson H, Kovats RS, McGregor GR, Stedman JR, Gibbs M, et al. 2005. The impactof the 2003 heatwave on mortality and hospital admissions in England. Health Stat. Q.2005:6–11
37. Kaiser R, Le Tertre A, Schwartz J, Gotway C, Daley R, Rubin CH. 2007. The effectof the 1995 heatwave in Chicago on all-cause and cause-specific mortality. Am. J. PublicHealth 97:S158–62
38. Kosatsky T, Litvak E, Analitis A, Katsouyanni K, Menne B, Penteli X. 2007. The inde-pendent and synergistic short-term effects of temperature and air pollution on health:review of the evidence. In Preparedness and Response to Heat-Waves in Europe: From Evidenceto Action, ed. B Menne, F Matthies. Copenhagen: World Health Org. In press
39. Katsouyanni K, Trichopoulos D, Zavitsanos X, Touloumi G. 1988. The 1987 Athensheatwave. Lancet 2:573
40. Katsouyanni K, Zmirou D, Spix C, Sunyer J, Schouten JP, et al. 1995. Short-term effectsof air pollution on health: a European approach using epidemiological time-series data.The APHEA project: background, objectives, design. Eur. Respir. J. 8:1030–38
41. Keatinge WR, Donaldson GC, Cordioli E, Martinelli M, Kunst AE, et al. 2000. Heat-related mortality in warm and cold regions of Europe: an observational study. Br. Med. J.321:670–73
42. Kilbourne EM. 1997. Heat waves and hot environments. In The Public Health Consequencesof Disasters, ed. E Noji, pp. 245–69. New York: Oxford Univ. Press
9.12 Kovats · Hajat
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
43. Kilbourne EM. 1992. Illness due to thermal extremes. In Public Health and PreventativeMedicine, ed. JM Last, RB Wallace, pp. 491–501. Norwalk, CT: Appleton Lang
44. Kilbourne EM, Choi K, Jones TS, Thacker SB. 1982. Risk factors for heatstroke. Acase-control study. JAMA 247:3332–36
45. Kirch W, Menne B, Bertollini R, eds. 2005. Extreme Weather Events and Public HealthResponses. Berlin: Springer-Verlag
46. Klinenberg E. 2002. Heat Wave: A Social Autopsy of Disaster in Chicago. Chicago: Univ.Chicago Press
47. Koppe C, Jendritzky G, Kovats RS, Menne B. 2003. Heatwaves: Impacts and Responses.Copenhagen: WHO
48. Kosatsky T. 2005. The 2003 European heat waves. Eurosurveillance 10:148–4949. Kosatsky T, King N, Henry B. 2005. How Toronto and Montreal (Canada) respond to
heat. See Ref. 45, pp. 167–7150. Kovats RS, Ebi KL. 2006. Heatwaves and public health in Europe. Eur. J. Public Health
16:592–9951. Kovats RS, Hajat S, Wilkinson P. 2004. Contrasting patterns of mortality and hospital
admissions during heatwaves in London, UK. Occup. Environ. Med. 61:893–9852. Kovats RS, Jendritzky G. 2005. Heat waves and human health. In Climate Change and
Adaptation Strategies for Human Health, ed. B Menne, KL Ebi, pp. 63–98. Darmstadt:Steinkopff Verlag
53. Kunkel KE, Changnon SA, Reinke BC, Arritt RW. 1996. The July 1995 heatwave inthe midwest: a climatic perspective of critical weather factors. Bull. Am. Meteorol. Soc.77:1507–18
54. Lecomte D, de Penanster D. 2004. People living in Paris, dead during the August 2003heatwave and examined in Medicolegal Institute. Bull. Acad. Natl. Med. 188:459–69
55. Leonardi G, Hajat S, Kovats RS, Smith GE, Cooper D, Gerard E. 2006. Syndromicsurveillance use to detect the early effects of heat-waves: an analysis of NHS Direct datain England. Prev. Med. 51:194–201
56. Le Tertre A, Lefranc A, Eilstein D, Declercq C, Medina S, et al. 2006. Impact of 2003heat wave on all cause mortality in 9 French cities. Epidemiology 17:75–79
57. Levy D, Lumley T, Sheppard L, Kaufman J, Checkoway H. 2001. Referent selection incase-crossover analyses of acute health effects of air pollution. Epidemiology 12:186–92
58. Lumley T, Levy D. 2000. Bias in the case-crossover design: implications for studies ofair pollution. Environmetrics 11:689–704
59. Luterbacher J, Dietrich D, Xoplaki E, Grosjean M, Wanner H. 2004. European sea-sonal and annual temperature variability, trends and extremes since 1500. Microb. Ecol.303:1499–503
60. MacFarlane A, Waller RE. 1976. Short-term increases in mortality during heatwaves.Nature 264:434–36
61. Martınez-Navarro F, Simon-Soria F, Lopez-Abente G. 2004. Valoracion del impacto dela ola de calor del verano de 2003 sobre la mortalidad. Evaluation of the impact of theheat wave in the summer of 2003 on mortality. Gac-Sanit. 18:250–58
62. McMichael A, Wilkinson P, Kovats RS, Pattenden S, Hajat S, et al. 2007. Internationalstudy of temperature, heat and urban mortality: the ‘Isothurm’ Project. Int. J. Epidemiol.Manuscript submitted
63. Michelon T, Magne P, Simon-Delaville F. 2005. Lessons of the 2003 heatwave in Franceand action taken to limit the effects of future heat waves. See Ref. 45, pp. 131–40
64. Michelozzi P, de Donato F, Bisanti L, Russo A, Cadum E, et al. 2005. The impact of thesummer 2003 heatwaves on mortality in four Italian cities. Eurosurveillance 10:161–65
www.annualreviews.org • Heat Stress and Public Health 9.13
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
65. Michelozzi P, DeSario M, Accetta G, DeDonato F, Kirchmayer U, et al. 2006. Temper-ature and summer mortality: geographical and temporal variations in four Italian cities.J. Epidemiol. Community Health 60:417–23
66. Michelozzi P, Kirchmayer U, Katsouyanni K, Biggeri A, McGregor GR, et al. 2007.Assessment and prevention of acute health effects of weather conditions in Eu-rope, the PHEWE project: background, objectives, design. Environ. Health. In press,doi:10.1186/1476-069X-6-12
67. Mirchandani HG, McDonald G, Hood IC, Fonseca C. 1996. Heat-related deaths inPhiladelphia—1993. Am. J. Forensic Med. Pathol. 17:106–8
68. Naughton MP, Henderson A, Mirabelli M, Kaiser R, Wilhelm JL, et al. 2002. Heatrelated mortality during a 1999 heatwave in Chicago. Am. J. Prev. Med. 22:221–27
69. O’Connor M, Cardinal M-E, Shaykewich J, Kosatsky T. 2007. Heat Health WarningWorkshop Rep., Oct. 5–6, 2006, Montreal, Quebec, Can. Agence de la sante et de servicessociaux de Montreal
70. Oke TR. 1995. Boundary Layer Climates. London: Methuen71. Oke TR. 1973. City size and the urban heat island. Atmos. Environ. 7:769–7972. O’Neill M. 2003. Air conditioning and heat-related health effects. Appl. Environ. Sci.
Public Health 1:9–1273. O’Neill M, Hajat S, Zanobetti A, Ramirez AM, Schwartz J. 2005. Impact of control for
air pollution and respiratory epidemics on the estimated associations of temperature anddaily mortality. Int. J. Biometeorol. 50:121–29
74. Paixao E, Nogueira PJ. 2002. Estudio de onda de la calor deo Julho de 1991 em Portugal:efeitos na mortalidade. Lisbon: Obs. Nac. Saude (ONSA)
75. Palecki MA, Changnon SA, Kunkel KE. 2001. The nature and impacts of the July 1999heatwave in the midwestern United States: learning from the lessons of 1995. Bull. Am.Meteorol. Soc. 82:1353–67
76. Panagiotakos DB, Chrysohou C, Pitsavos C. 2004. Climatological variations in dailyhospital admissions for acute coronary syndromes. Int. J. Cardiol. 94:229–33
77. Parsons KC. 2003. Human Thermal Environments: The Effects of Hot, Moderate, and ColdEnvironments on Human Health, Comfort, and Performance. London: Taylor & Francis
78. Pascal L, Nicolau J, Ledrans M. 2005. Evaluation de l’Impact de la Vague de Chaleur de l’Etesur la Morbidite Hospitaliere Infantile. Rapport d’etude. Paris: Inst. Veille Sanitaire
79. Quayle R, Doehring F. 1981. Heat stress: a comparison of indices. Weatherwise 34:120–2480. Roaf S, Crichton D, Nicol F. 2005. Adapting Buildings and Cities to Climate Change.
London: Archit. Press81. Samet JM, Dominici F, Zeger SL, Schwartz J, Dockery DW. 2000. The National Mor-
bidity, Mortality, and Air Pollution Study. Part I: Methods and methodologic issues. Res.Rep. Health Eff. Inst. Rep. 94: 5–14
82. Schwartz J. 2005. Who is sensitive to extremes of temperature? A case-only analysis.Epidemiology 16:67–72
83. Schwartz J, Samet JM, Patz JA. 2004. The effects of temperature and humidity on hospitaladmissions for heart disease. Epidemiology 15:755–61
84. Semenza JC, McCullough JE, Flanders WD, McGeehin MA, Lumpkin JR. 1999. Excesshospital admissions during July 1995 heat wave in Chicago. Am. J. Prev. Med. 16:269–77
85. Semenza JC, Rubin CH, Falter KH, Selanikio JD, Flanders WD, et al. 1996. Heat-relateddeaths during the July 1995 heat wave in Chicago. N. Engl. J. Med. 335:84–90
86. Senat. 2004. La France et les Francais face a la canicule: les lecons d’une crise. Rapportd’informatoin no. 195 (2003–2004) de Mme Letard, MM Flandre, S Lepeltier, fait au nom dela mission commune d’information du Senat, depose le 3 Fevrier 2004. Paris, Fr.
9.14 Kovats · Hajat
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
ANRV337-PU29-09 ARI 29 October 2007 17:22
87. Sheridan S. 2003. Heat, mortality and level of urbanisation. Clim. Res. 24:255–6588. Smoyer-Tomic KE, Rainham DGC. 2001. Beating the heat: development and evaluation
of a Canadian Hot Weather Health Response plan. Environ. Health Perspect. 109:1241–4789. Stafoggia M, Forastiere F, Agostini D, Biggeri A, Bisanti L, et al. 2006. Vulnerability to
heat-related mortality: a multi-city population based case-crossover analysis. Epidemiology17:315–23
90. Steadman RG. 1984. A universal scale of apparent temperature. J. Clim. Appl. Meteorol.23:1674–87
91. Stott PA, Stone DA, Allen MR. 2004. Human contribution to the European heatwave of2003. Nature 432:610–14
92. Sullivan-Bolyai JZ, Lumish RM, Smith EW, Howell JT, Bregman DJ, et al. 1979. Hy-perpyrexia due to air-conditioning failure in a nursing home. Public Health Rep. 94:466–70
93. Thomas ND, Soliman H. 2002. Preventable tragedies—heat disaster and the elderly.J. Gerontol. Soc. Work 38:53–66
94. Vandentorren S, Bretin P, Zeghnoun A, Mandereau-Bruno L, Croisier A, et al. 2006.August 2003 heat wave in France: risk factors for death of elderly people living at home.Eur. J. Public Health 16:583–91
95. Vandentorren S, Suzan F, Medina S, Pascal M, Maulpoix A, et al. 2004. Mortality in 13French cities during the August 2003 heat wave. Am. J. Public Health 94:1518–20
96. Vanhems P, Gambotti L, Fabry J. 2003. Excess rate of in-hospital death in Lyons, France,during the August 2003 heat wave. N. Engl J. Med. 348:2077–78
97. Wallace RF, Kriebel D, Punnett L, Wegmann DH, Amoroso PJ. 2007. Prior heat illnesshospitalization and risk of early death. Environ. Res. 104:290–95
98. Watkins R, Palmer J, Kolkotroni M, Littlefair P. 2002. The London Heat Island—surfaceand air temperature measurements in a park and street gorges. ASHRAE Trans. 108:419–27
99. Welty LJ, Zeger S. 2005. Are the acute effects of particulate matter on mortality in theNational Morbidity, Mortality and Air Pollution Study the result of inadequate controlfor weather and season? A sensitivity analysis using flexible distributed lag models. Am.J. Epidemiol. 162:80–88
100. Whitman S, Good G, Donoghue ER, Benbow N, Shou W, Mou S. 1997. Mortality inChicago attributed to the July 1995 heat wave. Am. J. Public Health 87:1515–18
101. Zeger S, Irizarry R, Peng RD. 2006. On time series analysis of public health and biomed-ical data. Annu. Rev. Public Health 27:57–79
www.annualreviews.org • Heat Stress and Public Health 9.15
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.
www.annualreviews.org ● Heat Stress and Public Health C-1
Figure 1
Daily mortality in Greater London, 2003, by age group.
050
100
150
200
Jan. 1, 2003 April 1, 2003 July 1, 2003 Oct. 1, 2003 Jan. 1, 2004date
age >75 age 65–74age 15–64 age 0–14
Kovats.qxd 10/29/2007 05:29 PM Page C-1
Ann
u. R
ev. P
ublic
. Hea
lth. 2
008.
29. D
ownl
oade
d fr
om a
rjou
rnal
s.an
nual
revi
ews.
org
by L
ondo
n Sc
hool
of
Hyg
iene
& T
ropi
cal M
edic
ine
on 0
2/14
/08.
For
per
sona
l use
onl
y.